Owing to recent taxonomic changes to the Nannoniscidae Hansen, the concepts of the taxa within the family require clarification. Specimens of Nannoniscus oblongus Sars, putatively those examined by G.O. Sars, and specimens from Hjeltefjord, Norway were illustrated to clarify the concept of the type genus of the Nannoniscidae. Specimens used by Siebenaller and Hessler (1981) and several other recently-described taxa were evaluated. Standard views are argued to provide more consistent illustrations of morphology. The somite articulations of the posterior body were found to be variable and often inaccurately illustrated feature in nannoniscid taxonomy; this character complex is therefore unreliable for taxonomic concepts in the family. In replacement, new characters that distinguish this family from the Desmosomatidae Sars are described. These include the proximal segmentation of the antennal flagellum, a subdistal dorsal tooth on the left mandible incisor process and ventral pereonal insertions of the coxae. The composition and classification of the family is adjusted using this new information. Subfamilies recently proposed for the Nannoniscidae by George (2001) based on somite articulations are rejected. Diagnoses of several genera, including Saetoniscus Brandt, were reconsidered using this new information. This latter genus is found to be indistinguishable from Nannoniscus Sars and is placed in junior synonymy. A new diagnosis and a new key to the genera of the Nannoniscidae use the new character information, omitting somite articulations as a primary descriptor. New diagnoses for Nannoniscus Sars and Nannonisconus Schultz, and revised compositions for these genera are proposed. Nannoniscus intermedius Siebenaller & Hessler is transferred to Nannonisconus. Rapaniscus Siebenaller & Hessler is diagnosed and pereopods I–II of its type species, R. dewdneyi are illustrated.